Understanding Line Array Systems
By John Murray
Behind the buzz, there are a lot of factors at work in line arrays. An explanation and comparison of current models.
A LITTLE HISTORY
Line arrays have been around for over a half of a century as column speakers, and other than those made by Rudy Bozak here in the US, most were voice-range only. Their application was generally for highly reverberant spaces, where a narrow vertical dispersion avoided exciting the reverberant field, provided a higher Q (narrower dispersion pattern) and, as a result, improved intelligibility of the spoken word.
Never losing popularity in Europe as they did in America, it’s no wonder that L-Acoustics V-DOSC loudspeakers from France were the first to show the concert sound world that more level and smoother frequency response can come from fewer drivers in a line array. After everyone realized that for a given listening area, the drivers have no destructive interference in the horizontal plane and combine mostly inphase in the vertical plane, the race was on.
Basically, a line of sources will create a wavefront of sound pressure that is loosely cylindrical in nature at a particular range of wavelengths (frequencies). Its idealized shape is actually more like a section of a cake, and the wavefront surface area, as it expands only in the horizontal plane, doubles in area for every doubling of distance. This equates to a 3dB SPL loss of level for every doubling of distance.
An idealized point source, imperfectly represented by a loudspeaker or nonlinear cluster of loudspeakers, radiates in a spherical waveform rather than cylindrical. This wavefront expands to four times the area with each doubling of distance, which equates to a 6dB SPL loss for every doubling of distance. This is commonly known as the inverse-square law, and it applies to all point-source radiant energy. Hence the big advantage for a line array is that for a given number of transducers, the resulting long throw level can be much greater than for a non-line array, or point-source, loudspeaker system.
This is the term applied to the dispersion pattern, or response balloon of a line array. It simply means that when you stack a bunch of loudspeakers, the vertical dispersion angle decreases because the individual drivers are outof- phase with each other at positions off-axis in the vertical plane. The taller the stack is, the narrower the vertical dispersion will be and the higher the sensitivity will be on-axis. In the horizontal plane, an array of like drivers will have the same polar pattern as a single driver. Some believe that the horizontal pattern is wider than for a single driver, but they are mistaken, likely fooled by the fact that the level is louder off to the side due to the higher sensitivity of multiple drivers. However, the actual polar pattern remains the same as for a single driver.
In addition to the narrowing vertical coverage angles, the array length also determines what wavelengths will be affected by this narrowing of dispersion. The longer the array, the lower in frequency (longer in wavelength) the pattern control will occur.
There is a limit to the 3dB per doubling loss, and it’s at this point where the array is far enough away to appear to be more of a point source and its level begins to attenuate according to the inverse-square law at 6dB per doubling of distance. The transition between these two regions is known as the critical distance for the line array. The region closer than critical distance, and the region beyond it, is termed as the Fresnel and Fraunhofer regions, respectively, so named by Christian Heil of L-Acoustics. Unless you’re a true math dweeb, near-field region and far-field region roll off the tongue a bit easier.
The critical distance for a given line array length varies inversely with wavelength (frequency). This was also discussed in depth in the last issue. Shorter wavelengths (higher frequencies) have much farther critical distances than longer wavelengths (lower frequencies). In theory this means, at greater distances, a line array will maintain more high-frequency content than low. However, air attenuation of the highs will counteract this characteristic.
Articulated is the ten dollar term for curved. This describes the very-popular J-Array shape that most manufacturers currently offer, save one. To date, the Duran Audio Intellivox system is the only line array that covers from extreme near-field to far-field seating with a straight-line dead-hang approach. (Talking about articulated arrays with your clients is what gets your day rate increased and your job title changed from “sound tech” to “audio engineer.”)
This is also a term for curved arrays of a particular type. Spiral arrays describe a curve that is increasing in the rotational angle from one end to the other, just as the common J-Array does from top to bottom.
ARITHMETIC SPIRAL ARRAYS
Mark Ureda, consultant to JBL, mathematically determined that spiral arrays that increase their angle of curvature in even increments perform better. For example, at the top of a line array, the splay between cabinets is 0 degrees. Going down the array, the element boxes are successively splayed at 1 degree, 2 degrees, 3 degrees, etc. Or it could go in increments of 2 degrees (i.e.: 2 degrees, 4 degrees, 6 degrees, etc.). These are arithmetically increasing spiral arrays.
Lobes describe all the acoustical energy that emanates from a loudspeaker or group of loudspeakers. The specified coverage angle of a horn is its main lobe. Spurious lobes are those that emanate out in a non-useful direction from the source.
Much ado has been made about lobe steering. Visions come to mind of FOH guys moving loudspeaker coverage around with a joystick. Lobe steering is generally done by incrementally delaying drivers in a line array. This can only be done when the sources, (the drivers), are about 1/2 wavelength apart for a given frequency, and only in the direction of the line array’s axis. For typical live sound HF drivers with a 9-inch diameter, this means that they cannot be positioned close enough together to steer anything above 750 Hz. However, using adaptive apertures to mimic a long line of smaller sources enables some steering at shorter wavelengths.
Side lobes are artifacts of line arrays. They are called side lobes but actually emanate from the ends of the array, at the top and bottom, as a typical line array is viewed in use. They are caused by the individual elements being in-phase at a particular angle and wavelength at some off-axis position from the array’s main lobe. It is possible to eliminate side lobes, but there are limits and consequences to side-lobe elimination in line arrays.
GRADIENT SIDE LOBES
This is a synonymous term for side lobes. Gradient describes how these lobes occur at particular angles or grades with respect to the line array’s orientation. Professional progress terminology tip: use gradient side lobes rather than side lobes in your technospeak. Chicks dig it.
Another of the fundamental parameters of line arrays is the spacing between individual elements. The accepted limit is that for good line array behavior, the sources should be no more than 1/2 wavelength apart for a given frequency. This means that loudspeakers reproducing longer wavelengths can be spaced farther apart without any deterioration in performance. But since 1/2 wavelength at 15 kHz is just under one-half of an inch, HF devices can never be close enough. One manufacturer maintains that because of this, line arrays do not really work at very high frequencies. However, I disagree, because even at very short wavelengths, the 3dB loss per doubling of distance still holds true, and this is what defines the line array effect. (In my humble opinion.) What does result from driver spacing of more than 1/2 wavelength is more pronounced gradient side lobing.
LOGARITHMIC DRIVER SPACING
Duran’s Intellivox Series line array loudspeakers employ the logarithmic driver spacing technique. This provides denser driver spacing at short wavelengths and economizes on the number of drivers needed for longer wavelengths by spacing them in larger and larger logarithmic increments.
Isophasic aperture is my current favorite high-tech term. It describes the phase characteristic of the slot that loads the horn bell of some line array box HF sections. The perfect line array driver, particularly for very short wavelengths, is a ribbon driver like those used by SLS Loudspeakers. Compression drivers are more rugged and capable of higher output levels than a ribbon driver, but they do not have a linear phase signal at the mouth of a horn.
Perhaps the most interesting technique for an isophasic device is the patented mid-high frequency aperture by Adamson. It employs the longer path length method, and utilizes directional vanes to prevent excess vertical dispersion as well. This approach is used for both the highand mid-frequency sections of their line array systems. The mid-frequency energy exits via two vertical slots on either side of the high-frequency exit slot. The paths of the mid-frequencies curve around the HF chamber housing. All slots are isophasic.
With the slots of the MF section on each side of the HF slot, diffractional problems of each slot on the other could be very problematic. However, Brock Adamson came up with a unique solution: overlapping the crossover points between the mids and highs. This provides in-phase pressure fronts from the other slots to prevent diffractional interference in the frequency range where it would be a problem.
The term “tapering” is also commonly called “shading.” They are essentially interchangeable. One of the first tricks used to take advantage of the line array effect was frequency tapering. My earliest exposure to this technique was the Electro-Voice LR-4B column speaker. For low/mids, it used 6-inch by 9-inch cone drivers that had lowpass filters at successively lower frequencies for speakers placed farther out to the ends of the column. This resulted in a longer column at longer wavelengths and a shorter column at shorter wavelengths, producing a similar dispersion pattern and critical distance for all frequencies, which in turns provides a more balanced frequency response at all listening distances.
Another tapering/shading technique is amplitude shading. This is used in many current line array products to accomplish front fill coverage where the bottom hook of a J-Array covers the extreme near-field listeners. This technique is simply lowering the volume of the loudspeakers covering the nearfield seating with respect to the longthrow loudspeakers higher in the array.
Some line array systems offer more than one choice for vertical dispersion of the individual box elements in the array. They do this as a solution to cover the near-field and extreme nearfield seating in most venues. EAW has gone one step further by offering two different models, matching the vertical dispersion and output level so that the drivers produce equal mouth SPL throughout the array. They avoid any amplitude shading for the drivers covering the closer listeners by increasing the coverage angle of those box elements. Why is it important to avoid amplitude shading?
According to David Gunness, EAW director of research and development, whenever two wave fronts with different pressures are combined, there will be a discontinuity at the juncture of the two. This discontinuity will be audible as though it were a separate, non-coherent source (delayed loudspeaker). The result is transient smear and uneven frequency response. Divergence shading provides a wave front whose curvature varies, but whose pressure magnitude does not. Therefore there is no introduced time smear to the signal.
HORIZONTALLY SYMMETRIC ARRAYS
The majority of available line array systems are horizontally symmetric. Ideally, each band pass is a 1/2 wavelength wide strip that runs the entire length of the array. The advantage is that it avoids horizontal lobbing at the crossover-frequency band. It also requires symmetric pairs of inner mid and outer LF drivers flanking the HF sophistic ribbon.
The drawback to this approach is that for the mid-drivers to be within 1/2 wavelength of each other, they must be incorporated into the bell of the HF horn. The normal 90-degree angle causes reflections between the MF drivers and the discontinuous horn walls cause HF problems as well.
HORIZONTALLY ASYMMETRIC ARRAYS
EV, Meyer (on their smaller system), and NEXO have opted for an asymmetric design. This approach avoided the mid-frequencies in the horn bell problem and contends with the horizontal lobbing at crossover problem inherent in asymmetric designs. Choose your poison.
CARDIOID AND HYPERCARDIOID LF SECTIONS
Line arrays have great directional control in the vertical axis. Subwoofer systems, by nature of the very long wavelengths involved, do not have any directional control unless arrayed. Even then, because of the omni-directional nature of each element in the array, there is no front-to-back directionality. This causes muddiness on stage and low-frequency feedback problems. Enter cardioid and hypercardioid low-frequency sections.
FIR-BASED VS. IIR-BASED DSP FILTERING
IIR (Infinite Impulse Response) filters in a DSP processor act just like analog crossover and equalization filters. Their amplitude and phase characteristics are in a fixed relationship. So much boost or cut produces an exact corresponding change to the phase response.
FIR (Finite Impulse Response) filters are able to manipulate phase independently of amplitude and correct for distance-related cancellations between drivers if each driver is under individual DSP control. Some systems, like Intellivox, employ separate DSP processing and amplification for each driver in the array. These types of systems will define the next big step forward in loudspeaker technology.
So, the next time you want to impress the ladies at the local hall, tell ‘em “We’re gonna hang a logarithmicspaced, articulated spiral array in a horizontally asymmetric configuration employing frequency tapering and divergence shading, which will include isophasic high-frequency and mid-frequency apertures, hyper-cardioid low-frequency transducer sections, is controlled by finite-impulse response filtering digital signal processing, and works well with a psychoacoustic infector.” You might just get lucky…
Live Sound Technical Editor John Murray is a 26-year industry veteran working for EV, Midas, Media Matrix? and TOA. John has presented two AES papers, chaired three Syn-Aud-Con workshops and is a member of the TEF Advisory Committee and ICIA adjunct faculty. If you have a question you’d like to ask John, e-mail him at firstname.lastname@example.org